Also, fldigi has worked well for me, perhaps it's worth exploring its
implementation...
Richard
--
This SF.net email is sponsored by Windows:
Build for Windows Store.
http://p.sf.net/sfu/windows-dev2dev__
On 06/17/2013 02:43 PM, David Rowe wrote:
> Hello Frank,
>
> Yes a few weeks ago I did some work on re-adding the Serial PTT dialog
> (as an option alongside Hamlib). I was also to add a patch for Macs and
> some way to remove the need for libctb (perhaps using native Com port
> calls for Win32).
On Mon, Jun 17, 2013 at 1:43 AM, David Rowe wrote:
> Hello Frank,
>
> Yes a few weeks ago I did some work on re-adding the Serial PTT dialog
> (as an option alongside Hamlib). I was also to add a patch for Macs and
> some way to remove the need for libctb (perhaps using native Com port
> calls f
Hello Frank,
Yes a few weeks ago I did some work on re-adding the Serial PTT dialog
(as an option alongside Hamlib). I was also to add a patch for Macs and
some way to remove the need for libctb (perhaps using native Com port
calls for Win32).
However my work and enthusiasm stalled when I coul
I finally got back to getting my build up and going on Windows -
successfully!
I finished off the autotools-based make only because that's what I had
started. I'll try cmake next time.
As it turns out, I really need serial PTT in order to use FreeDV. The
vox/half duplex option never worked for
On Thu, Jun 6, 2013 at 3:42 PM, David Rowe wrote:
> > I'm assuming you've been able to replicate the problem with the same
> > svn revision comparing both the autotools and cmake builds?
>
> Last time I tried, yes. have you managed to replicate the problem?
I can build both the libraries but i
> David,
>
>
> I'm assuming you've been able to replicate the problem with the same
> svn revision comparing both the autotools and cmake builds?
>
>
> Thanks,
> Richard
Last time I tried, yes. have you managed to replicate the problem?
- David
> --
As an aside I imagine the carriers could be made a little closer (like
25%) without significant degradation in BER at our usual SNR operating
points. This could be easily tested with the Octave simulations.
- David
On Wed, 2013-06-05 at 13:37 -0700, Bruce Perens wrote:
> Yes, in theory you can d
On Thu, Jun 6, 2013 at 9:43 AM, Bruce Perens wrote:
> On 06/06/2013 07:40 AM, Richard Shaw wrote:
> >
> > The results are not quite the same using the c2enc/c2dec
> This is going to be a floating point optimization issue.
Using the floating point math GCC wiki[1] as a guide I have not found
an
On Thu, Jun 6, 2013 at 9:43 AM, Bruce Perens wrote:
> On 06/06/2013 07:40 AM, Richard Shaw wrote:
> >
> > The results are not quite the same using the c2enc/c2dec
> This is going to be a floating point optimization issue.
Do you know the relevant flags that could affect this?
Thanks,
Richard
On 06/06/2013 07:40 AM, Richard Shaw wrote:
>
> The results are not quite the same using the c2enc/c2dec
This is going to be a floating point optimization issue.
--
How ServiceNow helps IT people transform IT departments:
Forgot to mention, even with the difference I've had successful QSO's with
FreeDV, not that that means we shouldn't find the root cause, but it's a
data point...
Richard
--
How ServiceNow helps IT people transform IT depar
On Thu, Jun 6, 2013 at 9:29 AM, Bruce Perens wrote:
> On 06/05/2013 02:23 PM, David Rowe wrote:
> > Also ATM the cmake build of codec2-dev gives different results to
> autotools.
> Oops. What's the symptom?
The results are not quite the same using the c2enc/c2dec. I've gone over
the build flag
On 06/05/2013 02:23 PM, David Rowe wrote:
> Also ATM the cmake build of codec2-dev gives different results to autotools.
Oops. What's the symptom?
--
How ServiceNow helps IT people transform IT departments:
1. A cloud ser
Hi Frank,
I've built libctb OK under mingw/msys and have added notes to
README.Win32. However I only did it once, so yes it wouldbe great if
this initial build process was made really easy.
I have a patch from the guys who build FreeDV on the Mac that also
removes the need for libctb on Linux &
> As David just mentioned, the plan is to move away from libctb and use
> hamlib. The only problem I've found so far is that it seems to be only
> designed to use serial ports for CAT control. I would have thought it
> would have a simple serial module to do simple serial PTT keying which
> I nee
Sure.
-Frank
On Wed, Jun 5, 2013 at 4:39 PM, Bruce Perens wrote:
> Do please graph the error rate as David has done on his blog.
>
>
> On 06/05/2013 01:33 PM, Frank Palazzolo wrote:
>
>
> Hi Bruce,
>
> Actually, I almost added that one - we will certainly find out if it is
> an issue. :)
>
101m.pdf
>
> Mel
>
> - Original Message -
> From: "Bruce Perens"
> To:
> Sent: Wednesday, June 05, 2013 3:22 PM
> Subject: Re: [Freetel-codec2] Building on Windows
>
>
> > Add "4) Is the linearity of Amateur equipment intended for
Wednesday, June 05, 2013 3:22 PM
Subject: Re: [Freetel-codec2] Building on Windows
> Add "4) Is the linearity of Amateur equipment intended for SSB voice and
> lacking in headroom beyond its rated power really adequate for this
> modulation?"
>
> On 06/05/2013 01:16 PM,
On 06/05/2013 01:55 PM, Bruce Perens wrote:
The problem is that this has to avoid overlap in any of
/number-of-carriers/ * /number-of-symbols/ cases
Nope. It's a power function. /number-of-symbols-per-carrier/ raised to
the power of /number-of-carriers./
-
That is actually the content of some Patent ideas, that is,
randomizing the data to improve PAPR. I think David recently discussed
that in his blog.
Hey, maybe it's time for a new amplifier to go with the new CODEC :-)
--
On 06/05/2013 01:48 PM, Gullik Webjörn wrote:
Enough rambling, the point was that all these carriers "added up"
causing high PAPR , the solution was to "randomize"
the PHASE, and minimize "adding up". Perhaps something similar could
be done here, i.e. each carrier
has a certain phase offset th
Or you could squeeze in 50% more information in the same noise
bandwidth, which could be
used for FEC or somesuch, or something else *fun*..
Given that this implementation is heavily aimed at "existing HF gear",
it does not make much sense
going *much* below standard SSB filter bw. People will
A small idea here: I once saw an app note on a "tunnel information
system" designed for
the old MW AM band. The designer clocked a rom at a modulus of 9 khz (
channel spacing)
and produced a sum of sines, 500 khz to 1.5 Mhz, i.e. a carrier on every
possible AM frequency.
The entire waveform wa
Do please graph the error rate as David has done on his blog.
On 06/05/2013 01:33 PM, Frank Palazzolo wrote:
Hi Bruce,
Actually, I almost added that one - we will certainly find out if it
is an issue. :)
I think I've convinced myself that all multicarrier methods suffer
from this same issu
Yes, in theory you can do without the guard bands, which is worthy of
exploration. But you can slow down the symbol rate or use fewer
carriers, so you don't really have to transmit faster.
On 06/05/2013 01:32 PM, Gullik Webjörn wrote:
This is something I raised a year ago or so. I have been lea
Hi Bruce,
Actually, I almost added that one - we will certainly find out if it is an
issue. :)
I think I've convinced myself that all multicarrier methods suffer from
this same issue, and probably to the same degree. If you add a bunch of
sine waves, most of the time they average out, but once i
This is something I raised a year ago or so. I have been learned that
the very reason to use ofdm is
that the "next" carrier is located on the "previous" and the "nextnext"
NULL, i.e. there is a minimum
"right there", and that given imperfect filtering, with the integer
relation between carrier
Add "4) Is the linearity of Amateur equipment intended for SSB voice and
lacking in headroom beyond its rated power really adequate for this
modulation?"
On 06/05/2013 01:16 PM, Frank Palazzolo wrote:
>
>
> The questions I am trying to answer are:
> 1) Will the orthogonality of the OFDM carrier
Well, at this point I'm still experimenting, and I'm really focused on
providing the best possible HF modem for FreeDV. Of course, if I come up
with something that can be used elsewhere, all the better. Most of the
OFDM stuff has really complex synchronization and a lot of bits per symbol.
I'm s
>
> I noticed the cmakelists.txt files only later on, along with the sinking
> feeling that remembered hearing people talking about cmake at Dayton :)
>
As the author of all the CMake stuff, hopefully it was in a good light :)
LOL! I mean sinking feeling because I had gotten so far without it on
On Wed, Jun 5, 2013 at 11:32 AM, Frank Palazzolo wrote:
> Thanks Richard - I should have stated first that I was going the
> Mingw/MSYS route using configure and make, following the README files in
> the codec2-dev and fdmdv2 directories. They didn't mention ctb there and
> so it was a surprise.
Frank,
I have built libctb from source on Windows using MinGW and MSYS since the
beginning of the project. I have not tried it using Cmake. I believe that
there was some problem with the Makefile that was distributed by the
author, at least a year ago, but it was simple to fix.
I will have to r
Thanks Richard - I should have stated first that I was going the Mingw/MSYS
route using configure and make, following the README files in the
codec2-dev and fdmdv2 directories. They didn't mention ctb there and so it
was a surprise. I noticed the cmakelists.txt files only later on, along
with the
On Wed, Jun 5, 2013 at 9:11 AM, Frank Palazzolo wrote:
>
> So, I've done some experimenting with an OFDM modem implementation in
> Octave. I still think the idea has a lot of promise, so and I've decided
> it's time to try to build codec2 and FreeDV.
>
> Codec2-dev was no big deal, as it has no d
: [Freetel-codec2] Building on Windows
hello Frank,I am not a programmer bur very interested in a windows application
for the OFDM modem, not necesarly Codec2 integrated.Just a OFDM modem for SW
with an easy Gui in windows for data transmissions. Idealy I want be capable to
run ip traffic through
hello Frank,
I am not a programmer bur very interested in a windows application for the
OFDM modem, not necesarly Codec2 integrated.
Just a OFDM modem for SW with an easy Gui in windows for data
transmissions. Idealy I want be capable to run ip traffic through the
modem. Speed should be autobaud a
So, I've done some experimenting with an OFDM modem implementation in
Octave. I still think the idea has a lot of promise, so and I've decided
it's time to try to build codec2 and FreeDV.
Codec2-dev was no big deal, as it has no dependancies. I didn't have
trouble getting the mingw/MSYS toolchai
38 matches
Mail list logo
